A Pathway-Based Approach to Association Analysis in Autism

Background: Recent advances in high-throughput genotyping have made it possible to conduct large-scale genome-wide association studies (GWAS) at population level to identify gene variants and genes associated with risk for common human diseases and a variety of psychiatric disorders such as autism. Over, the last several years many gene variants and candidate genes associated with autism have been identified using GWAS. However, the full breadth of the goals of high-throughput genotyping and GWAS to dissect the genetic architecture of autism is rapidly running into several bottlenecks in translating findings and hypothesis from GWAS to clinical practice to improve human health. One of the more significant bottlenecks is the inability of current GWAS analytic techniques to identify causal pathways and to characterize the functions of identified gene variants and candidate genes. Objectives: The objective was to determine if variation associated with autism tends to aggregate or cluster in biological pathways and gene networks. Methods: We conducted a gene and pathway-based association analysis using information from 2 genome-wide association studies to identify pathways and model gene networks involved in autism. We hypothesized that gene variants associated with autism map to and destabilize multiple candidate genes interacting within pathways and gene networks. Results: Our results show that genes containing genetic variants associated with autism are functionally related and interact with each other and their downstream targets within pathways and gene networks. Using publicly available gene expression data set, we are validating the results to infer the causal association between gene expression and autism.Conclusions: Our analysis demonstrates that integrative genomics leveraging information from GWAS with pathway analysis provides a powerful unified approach to autism biomarker discovery.